docs: add snippet to evaluate ARIMA plus model in the Forecast a single time series with a univariate model tutorial (#1267)

* merge main

* details

* Update samples/snippets/create_single_timeseries_forecasting_model_test.py

Co-authored-by: Tim Sweña (Swast) <swast@google.com>

* expand ibis to bq type conversion

* added unit test for from_ibis

---------

Co-authored-by: Tim Sweña (Swast) <swast@google.com>

Author: rey-esp

samples/snippets/create_single_timeseries_forecasting_model_test.py

@@ -72,9 +72,27 @@ def test_create_single_timeseries() -> None:
     # Expected output:
     #       ar_coefficients   ma_coefficients   intercept_or_drift
     #   0	 [0.40944762]	   [-0.81168198]	      0.0
-
     # [END bigquery_dataframes_single_timeseries_forecasting_model_tutorial_coef]
+
+    # [START bigquery_dataframes_single_timeseries_forecasting_model_tutorial_evaluate]
+    # Evaluate the time series models by using the summary() function. The summary()
+    # function shows you the evaluation metrics of all the candidate models evaluated
+    # during the process of automatic hyperparameter tuning.
+    summary = model.summary(
+        show_all_candidate_models=True,
+    )
+    print(summary.peek())
+
+    # Expected output:
+    # row   non_seasonal_p	non_seasonal_d	non_seasonal_q	has_drift	log_likelihood	AIC	variance	seasonal_periods	has_holiday_effect	has_spikes_and_dips	has_step_changes	error_message
+    #  0	      0	              1	               3	      True	     -2464.255656	4938.511313	     42772.506055	        ['WEEKLY']	            False	        False	            True
+    #  1	      2	              1	               0	      False	     -2473.141651	4952.283303	     44942.416463	        ['WEEKLY']	            False	        False	            True
+    #  2	      1	              1	               0 	      False	     -2479.880885	4963.76177	     46642.953433	        ['WEEKLY']	            False	        False	            True
+    #  3	      0	              1	               1	      False	     -2470.632377	4945.264753	     44319.379307	        ['WEEKLY']	            False	        False	            True
+    #  4	      2	              1	               1	      True	     -2463.671247	4937.342493	     42633.299513	        ['WEEKLY']	            False	        False	            True
+    # [END bigquery_dataframes_single_timeseries_forecasting_model_tutorial_evaluate]
     assert coef is not None
+    assert summary is not None
     assert model is not None
     assert parsed_date is not None
     assert total_visits is not None

tests/unit/core/test_dtypes.py

@@ -12,6 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import bigframes_vendored.ibis.backends.bigquery.datatypes as ibis_bq_types
 import bigframes_vendored.ibis.expr.datatypes as ibis_dtypes
 import bigframes_vendored.ibis.expr.types as ibis_types
 import geopandas as gpd  # type: ignore
@@ -152,6 +153,15 @@ def test_ibis_dtype_to_arrow_dtype(ibis_dtype, arrow_dtype):
     assert result == arrow_dtype
 
 
+@pytest.mark.parametrize(
+    ("ibis_dtype", "bigquery_type"),
+    [(ibis_dtypes.String(), "STRING"), (ibis_dtypes.String(nullable=False), "STRING")],
+)
+def test_ibis_dtype_to_bigquery_type(ibis_dtype, bigquery_type):
+    result = ibis_bq_types.BigQueryType.from_ibis(ibis_dtype)
+    assert result == bigquery_type
+
+
 @pytest.mark.parametrize(
     ["bigframes_dtype", "ibis_dtype"],
     [

third_party/bigframes_vendored/ibis/backends/bigquery/datatypes.py

@@ -55,6 +55,8 @@ def from_ibis(cls, dtype: dt.DataType) -> str:
             return "INT64"
         elif dtype.is_binary():
             return "BYTES"
+        elif dtype.is_string():
+            return "STRING"
         elif dtype.is_date():
             return "DATE"
         elif dtype.is_timestamp():